From d3f007b7f63ce3ac724e0336e7435c9a9fc27eac Mon Sep 17 00:00:00 2001
From: Trianta <56975502+Trimutex@users.noreply.github.com>
Date: Fri, 8 Dec 2023 18:03:58 -0600
Subject: [PATCH] Added reading in text file and trie structure

---
 src/generator.cpp | 60 +++++++++++++++++++++++++++++++++++++++++++++++
 src/generator.hpp | 25 ++++++++++++++++++++
 src/main.cpp      |  2 ++
 3 files changed, 87 insertions(+)

diff --git a/src/generator.cpp b/src/generator.cpp
index d9edafa..ab8f2d1 100644
--- a/src/generator.cpp
+++ b/src/generator.cpp
@@ -1,6 +1,36 @@
 #include "generator.hpp"
+#include <fstream>
 #include <iostream>
 
+
+void Trie::insert(const std::deque<char>& currentKGram) {
+    TrieNode* current = root;
+
+    for (char ch : currentKGram) {
+        if (current->children.find(ch) == current->children.end()) {
+            current->children[ch] = new TrieNode();
+        } else { ++current->occurances; }
+
+        current = current->children[ch];
+    }
+
+    current->isEndOfWord = true;
+}
+
+bool Trie::search(const std::deque<char>& currentKGram) const {
+    TrieNode* current = root;
+
+    for (char ch : currentKGram) {
+        if (current->children.find(ch) == current->children.end()) {
+            return false;
+        }
+
+        current = current->children[ch];
+    }
+
+    return current->isEndOfWord;
+}
+
 void Generator::SetArguments(int argc, char* argv[]) {
     std::string tempStr;
     for (int i = 1; i < argc; i += 2) {
@@ -28,6 +58,36 @@ void Generator::SetArguments(int argc, char* argv[]) {
     if (!setup.isFileSet || !setup.isPrefixSet || !setup.isOutputSet) { PrintUsage(); }
 }
 
+void Generator::ReadFile(void) {
+    std::ifstream inputFile(setup.filename);
+    if (!inputFile.is_open()) {
+        std::cerr << "[ReadFile - Error] Could not open file: " << setup.filename << std::endl;
+        exit(1);
+    }
+    std::deque<char> currentKGram;
+    char tempChar;
+    // Read in first k-gram
+    {
+        std::string initializeKGram;
+        inputFile.get(&initializeKGram[0], setup.prefixLength);
+        for (char ch : initializeKGram) { currentKGram.emplace_back(ch); }
+        trie.insert(currentKGram);
+    }
+    // Read rest of file
+    while (inputFile.get(tempChar)) {
+        currentKGram.emplace_back(tempChar);
+        if (currentKGram.size() > setup.prefixLength) { currentKGram.pop_front(); }
+        trie.insert(currentKGram);
+    }
+}
+
+void Generator::GenerateOutput(void) {
+}
+
+char GenerateCharacter(void) {
+    return 'z';
+}
+
 void PrintUsage(void) {
     std::cout << "Usage:  markov -i input_file -k prefix_length -n output_length" << std::endl;
     std::cout << "        -i: Direct path to input file for basis" << std::endl;
diff --git a/src/generator.hpp b/src/generator.hpp
index a50313b..5c5e76f 100644
--- a/src/generator.hpp
+++ b/src/generator.hpp
@@ -1,6 +1,8 @@
 #ifndef GENERATOR_HPP
 #define GENERATOR_HPP
 
+#include <deque>
+#include <unordered_map>
 #include <string>
 
 void PrintUsage(void);
@@ -14,13 +16,36 @@ struct ArgumentList {
     bool isOutputSet = false;
 };
 
+class TrieNode {
+public:
+    std::unordered_map<char, TrieNode*> children;
+    int occurances = 1;
+    bool isEndOfWord;
+
+    TrieNode() : isEndOfWord(false) {}
+};
+
+class Trie {
+public:
+    Trie() : root(new TrieNode()) {}
+    void insert(const std::deque<char>& currentKGram);
+    bool search(const std::deque<char>& currentKGram) const;
+private:
+    TrieNode* root;
+};
+
+
 struct Generator {
 public:
     Generator(void) = default;
     ~Generator(void) = default;
     void SetArguments(int argc, char* argv[]);
+    void ReadFile(void);
+    void GenerateOutput(void);
 private:
+    char GenerateCharacter(void);
     ArgumentList setup;
+    Trie trie;
 };
 
 #endif // !GENERATOR_HPP
diff --git a/src/main.cpp b/src/main.cpp
index 6997419..56497de 100644
--- a/src/main.cpp
+++ b/src/main.cpp
@@ -7,5 +7,7 @@ int main(int argc, char* argv[]) {
     }
     Generator markovChain;
     markovChain.SetArguments(argc, argv);
+    markovChain.ReadFile();
+    markovChain.GenerateOutput();
     return 0;
 }